Real-time Odometry-less 3D LiDAR SLAM with Generalized ICP and Pose-Graph Optimization (Long Ver.)

A research project in progress in our lab. This project is going for development of real-time lidar based SLAM for smart mobility.
-----
Mobile Robotics Research Team (MR2T) @ AIST
Twitter: / mr2t_aist
Webpage: unit.aist.go.j...

[@The_Funkyman]

Very nice. The last minute looks even greater than the rest. I love to see the model moving thanks to loop closure.

[@guoweidu7614]

Hi ~ couple of questions,
What is the lidar used and is the method opensource?

[@user-nw1rh5ut1d]

Have you publish the paper about this method or open-source the code ?

[@zainmehdi8886]

Woww looks awesome. I am also starting out on my work for 3d lidar SLAM. I already have experience with 2D version. Would you be kind enough to share some details ? Any paper or relevant literature you followed ?

[@yuhunk8053]

What does 2d experience mean? Single-line or multi-line mapping? 3d slam You can refer to lego-loam or hdl-slam

[@yuhunk8053]

Did you use imu for attitude compensation?

[@shitengshi2983]

Nice work. I want to know what does 'odometry-less' mean? Does it mean that odometry is provided by other modules and lidar is only responsible for mapping?

[@rubensleite5838]

I guess this means the odometry comes from the sequential lidar point cloud registration

[@diodin8587]

It means there is no odometry(relative-pose) factor in the factor graph but directly using scan matching cost.@@rubensleite5838

[@Jasonreninsh]

is this the same tech as GEOSLAM?

[@vigigamehacker123]

at 7:40 it seems to correct itself, can anyone explain what happened?

[@romaing.1510]

You can search for "loop closure SLAM" to have some more details

[@hiabhiabhi]

Hi, impressive video and technology. Can you please send me a step by step tutorial of how to achieve this?